The Resume & Applying

What actually happens to your resume

Before you write a word, understand the gauntlet your resume runs. When you click "apply," it doesn't go straight to a hiring manager who reads it carefully over coffee. It hits two filters in sequence, and you have to pass both:

The first reader is software — an applicant tracking system (ATS). It parses your file into fields, scans for keywords from the job description, and either filters you out or ranks you for a human. The second reader is a busy recruiter who spends roughly six to eight seconds on the first pass, scanning for: the right title, the right skills, a few numbers that signal impact, and recency. If either filter trips, you're out — no matter how good you'd be at the job.

Resume structure for a data engineer

Keep it to one page. You're early-career or career-changing; a second page dilutes your strongest material and signals you don't know what matters. Use this order, and notice how it differs for a career-changer:

CONTACT       Name · City · email · phone · LinkedIn URL · GitHub URL
              (no full street address; no photo)

SUMMARY       2-3 lines. Who you are + what you do + what you're targeting.
              "Career-changing data engineer with a production-grade ELT
               platform (dbt, Airflow, Snowflake). Former [X] who now builds
               reliable, tested data pipelines."

SKILLS        The literal tools, grouped. THIS is where ATS keywords live.
              Languages: SQL, Python
              Transform/Orchestrate: dbt, Airflow, Dagster
              Cloud/Warehouse: AWS, Snowflake, BigQuery
              Infra: Docker, Git, Linux, Terraform

PROJECTS      ← For a career-changer, put this HIGH (right after Skills).
              Your capstone goes here. This is your proof you can do the job.

EXPERIENCE    Your work history, with impact bullets. Even non-data jobs count
              if you frame the transferable parts (data, automation, scale).

EDUCATION     Degree(s), bootcamp/certs. One or two lines. Keep it last.

Why a dedicated Skills section matters so much: it's the densest, most scannable place for the exact tool names a recruiter and an ATS are hunting for. Don't make them infer "they probably know dbt" from a buried sentence — list it plainly. We'll come back to which keywords to include in the ATS and tailoring sections.

Impact-driven bullets

This is where most resumes are weakest and where you can stand out fastest. A weak bullet describes a task ("responsible for managing reports"). A strong bullet proves an outcome. Use this formula on every single bullet:

Strong verbs to open with: built, automated, designed, reduced, migrated, optimized, scaled, shipped, eliminated, cut. Avoid limp openers like "responsible for," "helped with," "worked on." Here's the transformation in practice:

Featuring your projects when you lack DE experience

Here is the career-changer's secret weapon. You may not have a "Data Engineer" job on your resume yet — but you have something most applicants don't: a real, working, documented platform from your capstone. Treat it exactly like a job. Give it a name, a one-line description of what it does, and impact bullets using the same formula.

For each project, answer three things: what it does, the stack, and what you personally handled. Here's the shape:

mini-GridDP — data platform for a GPU-rental marketplace    [GitHub] [demo]
Python · SQL · dbt · Dagster · DuckDB · Redpanda · Docker · GitHub Actions

• Built an end-to-end platform ingesting rentals from Postgres and ~[N] GPU
  telemetry events/day from a Redpanda stream into a DuckDB warehouse,
  modelled bronze → silver → gold by a [N]-model dbt project with tests.
• Orchestrated nightly runs in Dagster with retries and failure alerts; added
  data-quality tests that block bad data from reaching the Metabase dashboard.
• Containerized the full stack with Docker Compose and wired CI to build and
  test every pull request, so the whole platform spins up from one command.

Pair every project with a working GitHub link and, if you can, a short README or demo. The link turns a claim into evidence — and a clickable repo is exactly what a curious hiring manager wants on that six-second pass.

Beating the ATS

The applicant tracking system parses your file and matches it against the job description. Most early rejections aren't human decisions — they're parsing failures and missing keywords. You beat the ATS by being boring and explicit:

Tailoring per role

A single generic resume is the slowest path to a job. You don't need a from-scratch rewrite each time — but you should tailor at least the summary and skills to each posting. Read the JD, note its top 5–8 hard skills and the phrases it repeats, and adjust:

• Summary: echo the role's framing. If they want someone "building reliable batch pipelines on AWS," your summary should say roughly that — in your own words, truthfully.
• Skills: reorder so the tools they emphasize appear first, and make sure every JD keyword you legitimately have is present.
• Bullets: if you have several, surface the ones most relevant to this role toward the top.

Application strategy

Where you apply matters more than how many times you apply. The single highest-yield channel, by a wide margin, is a referral — a current employee passing your resume to the recruiter. Referrals skip much of the ATS lottery and arrive with built-in credibility. Here's how channels stack up:

How to get referrals (you have more reach than you think):

• List people you know — former colleagues, classmates, bootcamp peers, friends — and check who works at companies you'd like.
• Reconnect before you ask. A genuine catch-up beats a cold "can you refer me."
• Make the ask easy: name the specific role and link, share your capstone, and offer your tailored resume. "Would you be comfortable referring me for this role? Here's the link and my resume — no worries either way."
• Use LinkedIn (Chapter 02) to find second-degree connections at target companies, and engage authentically before reaching out.

✓ Check yourself

Answer from memory before opening each one.

1. A posting asks for "Airflow" and "dbt". Your resume says you "built orchestrated transformation workflows" and lists your tools as icons in a sidebar of a two-column template. You have used both tools and are rejected within the hour. What most likely happened?

   Show answer

   No human decided anything: the ATS parsed the file, could not read the icons, may have scrambled the two columns, and found neither exact keyword — a synonym like "orchestrated workflows" does not match "Airflow". Use a single-column, real-text PDF with standard headings, and put the JD's exact terms for skills you genuinely have in the Skills section. See Beating the ATS.

2. Once the resume reaches a recruiter, roughly how long is the first pass, what are they scanning for, and what does that imply about where your strongest material goes?

   Show answer

   About six to eight seconds, scanning for the right title, the right skills, a few numbers that signal impact, and recency. So the resume is a one-page highlight reel aimed at one job, with the best evidence at the top — not a complete record of everything you have done. See What actually happens to your resume.

3. You spent six years as an operations analyst and have never held a data engineering title. What order do the sections go in, why, and when does that order change?

   Show answer

   Contact, Summary, Skills, then Projects — with the capstone — above Experience, and Education last. The classic order would bury your best evidence under six years of non-data work; leading with Skills + Projects puts "can do DE work" inside the six-second look, and Experience then reinforces it through its transferable parts. Once you have a data job, Experience goes back above Projects. See Resume structure.

4. Your capstone entry reads "Worked on a data pipeline project using dbt and an orchestrator." You have no employer metrics to cite. Rewrite it with the chapter's formula — what can you legitimately quantify?

   Show answer

   Accomplished X by doing Y, measured by Z, opening with a strong verb — for example: "Built an end-to-end ELT platform ingesting ~2M rows/day, transformed by a 30-model dbt project with tests that block bad data from reaching dashboards." A project can still be measured in rows processed, model count, test coverage, runtime, or the manual effort it replaces, and an honest "~" estimate beats no number. See Impact-driven bullets and Featuring your projects.

5. To be safe against the keyword filter, a friend suggests pasting the whole job description into your resume footer in white 1pt text. Separately, the JD lists Kafka, which you have never used. What do you do about each?

   Show answer

   Neither goes on the resume. Hidden keyword-stuffing is caught by modern systems and by humans, and it ends the application; listing a tool you do not have is a lie that the interview exposes. The keyword test covers only hard skills you genuinely have — make each of those appear verbatim, and leave the rest off. See the keyword test.

6. Two applications this week: a cold apply through a job board to a large company, and a role at a 40-person company where a former colleague has offered to refer you. How much tailoring and cover-letter effort does each deserve, and what do you do after ten days of silence?

   Show answer

   Both get at least a tailored summary and reordered skills. A cover letter on the cold apply is rarely read, so skip it unless it is required; the referred, small-company role is exactly where a tight 3–4 sentence note tying your capstone to their problem can tip the balance — and make the referral easy by sending the role link and tailored resume. Log both in the funnel sheet and follow up politely after 7–10 days. See Tailoring per role and Application strategy.

1. Responsible for moving data from CSV files into a database.
2. Worked on reports to help the marketing team.
3. Used Python scripts to do some automation.

1. Built a Python loader that ingests 20+ daily CSV feeds into Postgres
   with validation, replacing a manual import that took ~5 hours/week.

2. Automated a weekly marketing-metrics pipeline in SQL + dbt, cutting
   report turnaround from 2 days to under 1 hour and removing 3 manual steps.

3. Automated 4 recurring data-prep tasks with Python scripts on a schedule,
   eliminating ~8 hours/week of manual work and reducing input errors to near zero.

Takeaway

• Your resume must pass two filters in order, an applicant tracking system that matches keywords and a recruiter who gives it about six seconds.
• Keep it to one page, and as a career-changer lead with Skills and Projects because they show data engineering work before your experience can.
• Write each bullet as what you accomplished, how you did it and a number that proves it, and use a real figure from your build because an invented one fails the first follow-up question.
• Treat the capstone like a job entry with a name, a stack, what you personally handled and a working GitHub link.
• Beat the parser by being plain: one column, standard headings, a text PDF, and the job description's own keywords copied verbatim.
• Tailor the summary and skills to each posting, and spend your effort on referrals, since ten tailored applications with referral attempts beat a hundred generic ones.

A strong resume earns the technical screen, and the next chapter drills the queries and coding problems that gate most data roles. → Interview Prep I — SQL & Coding